NEW TYPES OF LEAD TUNGSTATE CRYSTALS WITH HIGH LIGHT YIELD

Transcription

1 Frascati Physics Series Vol. XXI (2), pp IX INT. CONF. ON CALORIMETRY IN PART. PHYS. - Annecy, Oct. 9-14, 2 NEW TYPES OF LEAD TUNGSTATE CRYSTALS WITH HIGH LIGHT YIELD R.H. Mao, G.H. Ren, D.Z. Shen, Z.W. Yin Shanghai Institute of Ceramics, Shanghai 25, P.R.C. X.D. Qu, L.Y. Zhang, R.Y. Zhu California Institute of Technology, Pasadena, CA 91125, U.S.A S.P. Stoll, C.L. Woody Brookhaven National Laboratory, Upton, NY 11973, U.S.A ABSTRACT Because of their high stopping power and fast scintillation, lead tungstate crystals have attracted much attention in the high energy physics and nuclear physics communities. The use of lead tungstate, however, is limited by its low light output. An effort has been made at the Shanghai Institute of Ceramics to improve this. The results indicate that up to a factor of ten increase in the light output, mainly in the microsecond decay component, may be achieved. The X-ray diffraction pattern, photo luminescence spectrum, light output, decay kinetics and transmittance spectrum of new samples are presented. Longitudinal uniformity of a sample of 22 radiation lengths is studied. Possible applications for calorimetry in high energy and nuclear physics experiments are discussed.

2 IX Int. Conf. on Calorimetry in Part. Phys., Annecy, Oct. 9-14, Introduction In the last five years an extensive R&D program has been carried out by the Compact Muon Solenoid (CMS) experiment in developing lead tungstate (PbWO ) crystals to be used in the Large Hadron Collider (LHC). As a result of this development program, a total of 11.2 m large size (25 X ) PbWO crystals with fast scintillation light will be produced in Bogoroditsk Techno-Chemical Plant (BTCP) in Tulla, Russia, and in Shanghai Institute of Ceramics (SIC) in Shanghai, China. Because of this development, PbWO crystal is now a mature material in market with low cost. It, however, is interesting to note that the Y doped PbWO crystals chosen by CMS have limited light output, about 1 p.e/mev for full size samples measured with a photo multiplier (PMT) of bi-alkali cathode. This limits their application in areas other than high energy and nuclear physics. One approach to modify scintillation property of a material is altering crystal structure by changing growth parameter. Doping during crystal growth is another approach which may compensate structure defects, eliminate unwanted impurities and change scintillation properties. In developing BGO crystals for the L3 experiment, Eu doping was used at SIC to improve its radiation resistance 1. In developing CsI(Tl) crystals for the and BELLE experiments, a special scavenger was used at SIC to remove oxygen contamination 2. Pentavalent (Nb) doping in PbWO was first reported by Lecoq et al. to be effective in improving transmittance at 1 ppm level 3. Trivalent (La) doping was first reported by Kobayashi et al. to be effective in improving both the transmittance 4 and the radiation hardness 5. Consequent studies on doping with various ions, such as La, Lu, Gd, Y and Nb, at optimized level were reported to be effective in improving transmittance as well as radiation hardness 6 7. Early Glow Discharge Mass Spectroscopy (GDMS) analysis revealed that contaminations of certain cation, especially Mo, were responsible for the slow scintillation component in PbWO, as reported by Kobayashi et al. 8 and Zhu et al. 9. On the other hand, Mo doping introduces a significant fraction of the slow component and thus increases the light output in PbWO crystals. Following this line, PbWO samples doped with various dopant were grown and were found with significant increase of light yield In this paper we present scintillation and other optical properties of PbWO crystals doped with two special dopant A and B 1. It is found 1 Pending on patent application, the chemical nature of particular dopant is not released at present.

3 IX Int. Conf. on Calorimetry in Part. Phys., Annecy, Oct. 9-14, that light output of up to ten folds of that of the CMS Y doped PbWO crystal, mainly in the microsecond decay component, can be achieved. PbWO crystals of this type may find applications in high energy and nuclear physics experiments, such as crystal calorimeters in future electron linear colliders or in heavy ion colliders, where interaction cross-section allows an integration time of a few s. 2 Samples A total of ten samples, grown by a modified Bridgman method at SIC, were studied. Table 1 lists the dimension, dopant, peak of the photo luminescence and the sample delivery date. As a comparison, a standard CMS Y doped sample S762, which is a 23 cm long tapered from cm to cm, is also listed in this table. Two different dopants (A and B) were introduced at different levels in the melt during growth. Table 1: List of Samples Investigated in this Paper ID Dimension(cm) (nm) Date Samples Doped with Dopant A S25! #"$!%&"'! 57 4/23/1999 S27 )(#"+*)("$)( 57 4/26/2 Z9 )(#"+*!,"')( 57 4/26/2 Z14 )(#"+*-.! #"$)( 57 7/21/2 Z22 )(#"+*/)("$)( 57 7/21/2 Z23 )("$-1"')( 57 7/21/2 Z24 )("$2)(#"')( 57 9/2/2 Z25 )(#"+*)("$)( 57 9/2/2 Samples Doped with Dopant B Z2 )(#"+*435)("$)( 57 7/21/2 Z21 )(#"+*6(7!2#"$)( 57 7/21/2 A Standard CMS Y Doped Sample S762!&"'2)(8"'!/ 43 8/25/2 All samples have rectangular shape with all surfaces polished. No further treatment, other than simple cleaning with alcohol, was carried out before measurements.

4 IX Int. Conf. on Calorimetry in Part. Phys., Annecy, Oct. 9-14, Intensity (a.u.) (2) (22) (116) (224) (112) (4) (24) (312) θ (degree) Figure 1: A photo showing, from top to bottom, a standard CMS Y doped sample S762, an A doped sample Z9 and a B doped sample Z2. Figure 2: XRD patterns for samples Z24 and Z25. The as-grown samples are transparent, colourless without visible defects, such as cracking, inclusions, scattering particles and growth striation. Figure 1 is a photo showing, from top to bottom, a standard CMS Y doped sample S762, an A doped sample Z9 and a B doped sample Z2. Both A and B dopings do not change crystal structure. Figure 2 shows the X- ray diffraction (XRD) patterns measured at SIC. Two small blocks cut from samples Z24 and Z25 were grounded into powders, and the XRD spectra were taken by using a D/MAX-C diffractometer with a Cu target running at 4 kv, 3 A and 2 9 /min. Both patterns match very well with standard power diffraction data of pure sheelite structure 12. No other phase was observed. 3 Emission Photo luminescence spectrum was measured by using a Hitachi F-45 fluorescence spectrophotometer equipped with a red sensitive Hamamatsu PMT R928, which extends measurement range from 22 to 8 nm. Figure 3 shows a schematic of the setup, where the UV excitation light was shot to a bare surface of a sample and the photo luminescence, without passing through sample to avoid effect of self absorption, was measured by an R928 PMT through a monochromator. Figure 4 shows the photo luminescence spectra for an A doped sample Z24, a B doped sample Z2 and a

5 @ < IX Int. Conf. on Calorimetry in Part. Phys., Annecy, Oct. 9-14, standard CMS Y doped sample S762. These spectra were corrected by the monochromator efficiency and the PMT quantum efficiency. The vertical axis refers to photon numbers, and its scale is arbitrary. As shown in Figure 4, both A and B doped samples have similar photo luminescence peaked at 57 nm, while that from sample S762 is peaked at 43 nm. This means that the scintillation of these new types of PbWO crystals is in green, contrary to the blue of standard CMS Y doped PbWO crystals. 15 Watt Xenon Lamp Excitation Monochromator Reference Phototube Photons (arbitrary unit) = 14 = 12 = 1 < 8 ; 6 4? S762 Z24 Z2 2 Crystal Emission Monochromator PMT (R928) : > ; Wavelength (nm) Figure 3: A schematic of setup used to measure photo luminescence. Figure 4: Photo luminescence spectra for samples Z24, Z2 and S Light Output and Decay Kinetics The scintillation light output was measured by using a Hamamatsu PMT R259, which has a bialkali photo cathode and a quartz window. To study crystal uniformity measurements were carried out by coupling both the seed and the tail end of a sample to the PMT. A thin layer of Dow Corning 2 fluid, which has very good UV transmittance 13, was used between the sample and the PMT, while all other faces of the CB sample were wrapped with two layers of the Tyvek paper. A collimated A Cs source was used to excite the sample, shooting at the non coupling end of the sample. The scintillation light pulse collected by the PMT was integrated by a LeCroy 31 QVT in the Q mode. A series of 8 integration gates, ranging from 35 to 4 ns, was provided by a LeCroy 2323A programmable gate generator to measure the light output as

6 E F G H IX Int. Conf. on Calorimetry in Part. Phys., Annecy, Oct. 9-14, a function of integration time. The D -ray peak was determined by a simple Gaussian fit, and was used to determine photoelectron numbers by using the single photoelectron peak. The measured photoelectron numbers can be converted to the photon yield of the sample by taking into account the quantum efficiency of the photocathode and the efficiency of the light collection. L.Y. (p.e./mev) Data T Corrected Data Time (hrs) Entries Mean RMS L.Y. (Data) Entries Mean RMS E L.Y. (T-Corrected Data) Figure 5: Light output measured by a PMT in 9 hours (top), and its distribution without (left bottom) and with (right bottom) temperature corrections. 3 Pedestal S762 Z2 Z Counts ADC Channels CB Figure 6: A Cs spectra integrated in 2 s for samples Z24, Z2 and S762. All measurement were carried out at the room temperature of 2 9 C. Data were corrected by using the room temperature which has daily variations of up to.5 9 C, despite central air condition of entire laboratory building and individual temperature adjustment and feedback in the room where measurement was carried out. This variation is significant since the light output of PbWO crystals is known to have -2%/9 C at room temperature. The systematic uncertainty of light output measurement, including temperature variation and operation uncertainties caused by mounting samples to the PMT, was estimated to be about 1%. With temperature corrections this uncertainty is reduced to.8%. The top plot in Figure 5 shows the repeated measurement of the light output for a sample in 9 hours. Two bottom plots in Figure 5 show raw (left) and temperature corrected (right) light output data and corresponding Gaussian fit. A precision of 1 and.8% were achieved for the light output measurement without and

7 N Q N IX Int. Conf. on Calorimetry in Part. Phys., Annecy, Oct. 9-14, with temperature corrections respectively. CB Figure 6 shows A Cs spectra integrated in 2 s from an A doped sample Z24, a B doped sample Z2 and a standard Y doped CMS PbWO sample S762. Both A and B doped samples have significantly more light than CMS sample, and the A doped sample Z24 has more light than the B doped sample Z2, part of which can be explained by the small size of sample Z24. Table 2: Summary of PbWO Light Output (p.e/mev) Sample Gate width (ns) Fraction(%) ID , 2, IKJML4N OMP OMP JKJML4N S25R S25S S27R S27S Z9R Z9S Z23R Z23S Z24R Z24S Z2R Z2S Z21R Z21S S R represents sample s seed end coupled to the PMT. S represents sample s tail end coupled to the PMT. Table 2 lists the light output integrated in five different gate widths for all PbWO samples listed in Table 1. Also listed in the table is the ratio of light outputs between 5, 1 and 2, ns. Significant increase of light output, especially in slow scintillation component, is observed for samples doped with A and B. With light output measured as a function of integration time, the scintillation decay kinetics of

8 ] ^ ` IX Int. Conf. on Calorimetry in Part. Phys., Annecy, Oct. 9-14, the samples was determined. Figure 7 shows a comparison of light outputs, in photoelectron per MeV, as a function of the integration time for samples Z24, Z2 and S762. The A and B doped samples have significantly increased slow component. 1 V8 \ Z24 ].3 ].25 HAMAMATSU PMT R259 (# BA1983) Light Output (p.e./mev) U6 4 Z2 Quantum Efficiency ].2 ].15 ].1 SIC-S762 a Q E=13.7 ±.3% SIC-Z24 a Q E=5.3 ±.1% 2 T [ S762 T W X Y Z Time (ns) ].5 3 ^ Wavelength (nm) ` 65 7 Figure 7: The light output is shown as a function of integration time for samples Z24, Z2 and S762. Figure 8: Quantum efficiency of R259 PMT is ahown as function of wavelength together with emission spectra of A doped sample Z24 and Y doped sample S762. On the face value of Table 2 and Figure 7, the A or B doped samples may provide photo electron yield of up to 5 to 8 times of that of the Y doped CMS crystal. To convert the measured photoelectron yield to the photon yield we measured emission weight quantum efficiency of the Hamamatsu R259 PMT used in the measurement. Figure 8 shows distributions of the quantum efificiency of R259 PMT and corresponding emission spectra for a CMS choice of Y doped PbWO sample S762 and an A doped PbWO sample Z24. The corresponding emission weighted quantum efficiencies are (bdcefhg1i c )% and (j ckg1ieb % respectively for S762 and Z24. The PMT response thus has a factor 2.6 difference for these two types of crystals. Calculation by using emission of a B doped sample shows consistent result. The light output of A or B doped samples in photon/mev thus is 13 to 21 times of that of Y doped PbWO. This number, however, has not taken into account the difference of the light path or crystal size. Taking into account the difference of the light path, our estimation is

9 p m l p IX Int. Conf. on Calorimetry in Part. Phys., Annecy, Oct. 9-14, that up to a factor of ten increase in light output, mainly in sec decay component, is expected for the A and B doped samples as compared to that of the standard Y doped CMS sample. Reading Table 2 one also notices that the difference of the light output between the seed and the tail end coupled to the PMT for the same sample. This is caused by the variation of the emission spectrum along the crystal length and will be discussed in details in Section 6. 5 Longitudinal Transmittance Longitudinal transmittance was measured by using a Hitachi U-321 UV/visible spectrophotometer with double beam, double monochromator and a large sample compartment equipped with a custom Halon coated integrating sphere. The systematic uncertainty in repeated measurements of transmittance was approximately.3%. Figure 9 shows longitudinal transmittance spectra for an A doped sample Z24, a B doped sample Z21 and a CMS Y doped sample S762. While the transmittance of Z24 is the highest, that of S762 is the lowest. This is partly due to the difference of the path length: 3, 1 and 23 cm respectively for samples Z24, Z21 and S762. One also notes that transmittance of sample Z24 approaches the theoretical limit. Transmittance (%) 8 o6 n4 2 Z21 tz24 us762 q n r o s Wavelength (nm) Photons (arbitrary unit) w2 1 v 2 1 v 2 1 v Em: 43 nm Em: 56 nm Ex: 33 nm Ex: 33 nm ~ ~ Em: 56 nm Ex: 33 nm { Seed } Middle part x y z Wavelength (nm) Tail Figure 9: Longitudinal transmittance for samples Z24, Z21 and S762. Figure 1: Photo luminescence measured at three positions along the axis of sample Z9.

10 ƒ IX Int. Conf. on Calorimetry in Part. Phys., Annecy, Oct. 9-14, Longitudinal Uniformity One important technical issue for doping is the uniformity. Since the segregation coefficient of a dopant in PbWO crystals is usually not equal to one, the dopant tends to distribute not uniformly in the crystal. This would in turn cuase bad longitudinal uniformity for large size samples. This bad longitudinal uniformity, if beyond some limit, may affect calorimeter performance. Our measurement shows that both dopants A and B are not uniformly distributed in PbWO. Table 2 shows that all A doped samples provide significant more slow component when the seed end is coupled to the PMT, while it is the tail end coupled to the PMT for the B doped samples. This indicates that the dopant A is concentrated at the tail end, and dopant B is concentrated at the seed end. In other words, the segregation coefficient of dopant A in PbWO is less than one, and that of dopant B is larger than one. To study longtitudinal doping uniformity a sample Z9 was grown to 22 radiation length with dopant A in the melt. The spectra of the photo luminescence, decay kinetics and transverse transmittance along the longitudinal axis of the sample Z9 are shown in Figures 1, 11 and Z9 Seed end coupling From top to bottom: Tail to seed 8 ŽZ9 Œseed Light Output (p.e./mev) Transmittance (%) tail middle part Time (ns) ˆ Š Wavelength (nm) Figure 11: Decay kinetics measured at four positions along the axis of sample Z9. Figure 12: Transverse transmittance measured at three positions along the axis of sample Z9. Figure 1 shows the exitation (left) and photo luminescence (right) spectra mea-

11 IX Int. Conf. on Calorimetry in Part. Phys., Annecy, Oct. 9-14, sured at the seed, middle and tail end of sample Z9. While the emission has both blue and green component at the seed end, it is pure green at the tail end. This is consistent with the segregation coefficient of less than one for dopant A in PbWO. The same conclusion can be drawn in decay kinetics measurement. Figure 11 shows that the light from the seed part of the sample has very little slow component, and that from the tail end has significant slow component. The transverse transmittance can also be used to judge sample longitudinal uniformity. Figure 12 shows good transverse transmittance at the seed end degrades to the tail end of the sample. Although this kind of degradation is sort of expected in Brigman method since all melt goes to the ingot, leading to a relatively large fraction of impurities in the tail end, the difference in the transmittance can be further reduced by defining growth parameters. An alternative solution of this problem is to grow longer ingots and cut off a longer tail part, which, however, would increase the cost. Similarly, measurements on B doped sample confirms that the segregation coefficient of dopant B is larger than one, causing more green light and more slow component at the seed end. 7 Summary In the last two years SIC has made an effort in developing new types of PbWO crystals with high light yield. It is encouraging to find both dopant A and B are effecive in increasing PbWO light output, and up to ten folds of light output increase is observed for these samples as compared to that of the standard Y doped CMS sample. We, however, have not be able to observe a dopant which cause significantly more fast component. This increase of slow component is encouraging for users in high energy and nuclear physics field, but may still fall short for medical applications. Both A and B doppings cause bad longitudinal uniformity. One interesting approach thus is to double dope PbWO crystals with both A and B. Since these two dopants have similar function but with rather different segregation coefficients, it is hoped that they would compesate each other and make large size, longitudianally uniform crystals. When this is achieved, we will have a new type of PbWO crystals for high energy and nuclear physics community.

12 IX Int. Conf. on Calorimetry in Part. Phys., Annecy, Oct. 9-14, Acknowledgments Work at Caltech is supported in part by U.S. Department of Energy Grant No. DE- FG3-92-ER471, and at BNL in part by U.S. Department of Energy Contract No. DE-AC2-CH7616. References 1. Z.Y. Wei et al., Nucl. Instr. and Meth. A297, 163 (199). 2. R.Y. Zhu, IEEE Trans. Nucl. Sci. NS-44, 468 (1997). 3. P. Lecoq et al., Nucl. Instr. and Meth. A365, 291 (1995). 4. M. Kobayashi et al., Nucl. Instr. and Meth. A399, 261 (1997). 5. M. Kobayashi et al., Nucl. Instr. and Meth. A44, 19 (1998). 6. S. Baccaro et al., Phys. Stat. Sol. A164, R9 (1997). 7. E. Auffray et al., Nucl. Instr. and Meth. A42, 75 (1998). 8. M. Kobayashi et al., Nucl. Instr. and Meth. A373, 333 (1996). 9. R.Y. Zhu et al., Nucl. Instr. and Meth. A376, 319 (1996). 1. R.Y. Zhu, in Proceedings of NLC Detector Workshop, Keystone, Colorado, September 1998; R.Y. Zhu, in Proceedings of International Workshop on Linear Collider, Sitges, Barcelona, Spain, May 1999; and in Proceedings of VIII International Conference on Calorimetry in High Energy Physics, ed. G. Barreira et al., World Scientific, 226 (1999). 11. P. Lecoq et al., in Proceedings of IEEE NSS/MIC, Seattle, November 1999, and A. Annenkov et al., Nucl. Instr. and Meth. A45, 71 (2). 12. Powder diffraction data file No , JCPDS-ICDD, 161 Park Lane, Swarthmore, PA 1981, USA, 222 (1979). 13. R.Y. Zhu, Nucl. Instr. and Meth. A34, 442 (1994).